View Single Post
  #83   Report Post  
Old February 25th 16, 01:49 PM posted to uk.sci.weather
Eskimo Will Eskimo Will is offline
external usenet poster
 
First recorded activity by Weather-Banter: Oct 2011
Posts: 3,280
Default How are the mighty fallen!


"Norman Lynagh" wrote in message
...
On Thu, 25 Feb 2016 13:21:12 -0000, "JohnD" wrote:

"Desperate Dan" wrote in message
...

The only way to understand a site is to survey the site.


Sorry, but I don't really agree, or at least not completely. There's no
argument but that the gold standard is for a professional survey of each
and
every site contributing data (to WOW in this particular instance). But
that's not practicable nor is it every likely to happen - although I
wouldn't exclude some more informed type of self-reporting/certification
than exists at present, maybe even encourage well-briefed amateurs in the
area to visit and comment.

However, that doesn't mean that all data from sites that haven't been
individually surveyed is equally bad and unreliable. Modern statistical
methods are quite powerful in monitoring the deviation of actuals from
gridded data calculated from the overall observational data set and should
quickly be able to prioritise site that offer consistently accurate data,
especially if the self-certification for exposure can be factored in too.
No-one is making the argument that this is ever going to be as good as
data
from a well-exposed and professionally surveyed site, but I do strongly
suspect that this approach may well be more powerful than you might
imagine.
Again I'd suggest looking at what CWOP does for the US.

JGD


Perhaps I am missing something here but if the modelling is good enough to
determine which actuals are 'good' and which are 'bad' then surely the
modelling
doesn't need these actuals at all.


Data assimilation is about nudging the model to a solution that better fits
the observational data. It does need actuals to fine tune its analysis.
After all forecasting short term is basically an initial value problem, and
so if data assimilation is done every 3 or 6 hours then nudging is required.
When continuous assimilation comes onboard every piece of data will be vital
since the short time-scales will require a greater accuracy on spatial
scales to avoid numerical instability in the assimilation process.

Will
--
" Some sects believe that the world was created 5000 years ago. Another sect
believes that it was created in 1910 "
http://www.lyneside.demon.co.uk/Hayt...antage_Pro.htm
Will Hand (Haytor, Devon, 1017 feet asl)
---------------------------------------------